A Boosted Evolutionary Neural Architecture Search for Timeseries Forecasting with Application to South African COVID-19 Cases

International Journal of Online and Biomedical Engineering (iJOE), Journal Year: 2023, Volume and Issue: 19(14), P. 107 - 130

Published: Oct. 11, 2023

In recent years, there has been an increase in studies on time-series forecasting for the future occurrence of disease incidents. Improvements deep learning approaches offer techniques modelling long-term temporal relationships. Nonetheless, this design practice is rigorously painstaking, prone to errors, and requires human expertise. The advent feature enrichment with automatic architecture search typically optimises discovery new neural architectures applicable domains such as modelling. main methodological contribution study approach using feature-enriched filters evolutionary sequence-to-sequence gated recurrent units (GRU-Seq2Seq). This applied prediction daily cases coronavirus South Africa. highly pathogenic pandemic incident data was modelled filters, optimised hyper-parameter trials evolutional algorithm. proposed model benchmarked against ARIMA SARIMA. predicted trends 30, 60 90-day horizons evaluated them 7, 14 31 days. Simulation results demonstrate that observed case counts added optimisation improve performance accuracy. Generally, bFilter+GRU-Seq2Seq optimal configuration outperformed SARIMA lower error scores higher metrics, R2 score 7.48E-01 a 30-day forecast horizon.

Language: Английский

The Concept of Islamic Spiritualism in Dealing with Plague and Infectious Disease Outbreaks

CSR, sustainability, ethics & governance, Journal Year: 2024, Volume and Issue: unknown, P. 493 - 501

Published: Jan. 1, 2024

Language: Английский